(Text by QI Jifeng, jfqi@qdio.ac.cn)
Jifeng participated in the Asia Oceania Geosciences Society (AOGS) annual meeting and discussed AI Oceanography with scholars. Credit: Institute of Oceanology, Chinese Academy of Sciences
At present, the field of marine science has entered a new era of big data. It is believed that the collective volume of marine big data has reached a staggering exabyte (EB) level. This brings unprecedented opportunities and challenges to marine science. With the rapid development of artificial intelligence (AI) technology, how to efficiently use this technology to process and analyze massive amounts of ocean data has become a key issue. This is crucial for a deep understanding of the complex dynamics of the ocean system.
AI technology can assist scientists in more accurately predicting changes in the marine environment, such as ocean currents, temperature changes, and sea-level rise, thus providing more precise data support for climate change research. Consequently, delving into the application of AI in physical oceanography has sparked considerable interest in academic circles.
As a research scientist at the Institute of Oceanology, Chinese Academy of Sciences (IOCAS), my research focuses on AI oceanography and the variations in ocean water masses, thermohaline structures, and their climatic effects. In 2014, I obtained my Ph.D. from the University of Chinese Academy of Sciences, where my research centered on the variation of ocean water masses and their climatic effects, laying the groundwork for my future studies.
In recent years, my main research has been focused on AI-driven oceanographic studies. We've developed an AI-based model to estimate the ocean's three-dimensional temperature and salinity structures. This model, rooted in multi-source ocean satellite remote sensing data and deep learning techniques, can deftly map the connection between surface satellite data and the underlying thermohaline structures, thus providing accurate estimations of these subsurface features. My research also delves into using AI for estimating key oceanic layers, such as the mixed layer depth, thermocline, barrier layer thickness, and sonic layer depth, all yielding fruitful results. Moreover, AI's application in ocean forecasting has led to the creation of intelligent models for rapid prediction of marine environmental elements and phenomena, like marine heat waves, demonstrating AI's transformative capabilities in this domain. The above research results not only break through the limitations of traditional observation techniques and numerical models but also demonstrate the enormous potential and broad application prospects of machine learning, especially deep learning and meta-learning, in oceanographic research.
Further, my research extends to exploring the variability of ocean water masses, thermohaline structures, and their climatic effects, such as revealing the mechanism of salinity changes in the tropical Pacific and its relationship with El Nino-Southern Oscillation (ENSO); studying in depth the impact of different types of El Nino on the salinity of the tropical Pacific and the mechanism of ENSO's influence on the salinity of the South China Sea; and clarifying the main characteristics and mechanisms of low-frequency variations of the Southwest Pacific subtropical mode water.
AI, as a rapidly evolving technology, offers streamlined and direct approaches for studying complex marine phenomena. It serves as an effective, agile tool for comprehending intricate marine processes and extracting valuable insights from the expanse of oceanic data. Looking forward, my research will continue to focus on studies combining AI technology with traditional physical oceanography mechanisms, contributing to the development of AI oceanography.
(Editor: ZHANG Yiyi)